Prior reproduction and weather affect berry crops in central Ontario, Canada

Page created by Raymond Rose
 
CONTINUE READING
Popul Ecol (2012) 54:347–356
DOI 10.1007/s10144-011-0301-6

 ORIGINAL ARTICLE

Prior reproduction and weather affect berry crops in central
Ontario, Canada
Eric J. Howe • Martyn E. Obbard          •

Jeff Bowman

Received: 22 February 2011 / Accepted: 11 November 2011 / Published online: 13 December 2011
Ó The Society of Population Ecology and Springer 2011

Abstract Populations of many perennial plants intermit-             dynamics, particularly in boreal and subArctic environ-
tently produce large seed crops—a phenomenon referred to            ments where berry crops are important wildlife foods.
as mast seeding or masting. Masting may be a response to
spatially correlated environmental conditions (the Moran            Keywords Berry crops  Endozoochory  Masting 
effect), an adaptive reproductive strategy reflecting econ-         Moran effect  Pulsed resources  Seedfall
omies of scale, or a consequence of the internal resource
budgets of individual plants. Fruit production by endozo-
ochorous plants representing eight genera varied synchro-           Introduction
nously over much of central Ontario, Canada, 1998–2009.
We tested for effects of weather and prior reproduction on          Populations of many polycarpic plants intermittently pro-
fruit production by comparing AICc values among regres-             duce large seed crops (Herrera et al. 1998; Kelly and Sork
sion models fit to time series of fruit production scores           2002). This phenomenon is referred to as mast seeding or
and partitioning contributions by different predictors to           masting, and several explanations for it have been proposed
multiple R2 into independent and joint contributions. Fruit         (Kelly 1994; Isagi et al. 1997; Kelly and Sork 2002). Seed
production by mountain ash (Sorbus spp.), juneberry                 crops could simply vary in response to spatially correlated
(Amelanchier spp.), dogwoods (Cornus spp.), nannyberry              environmental conditions (referred to as resource match-
(Viburnum lentago), and possibly cherries (Prunus spp.)             ing, climatic forcing, or the Moran effect; Royama 1992).
was inversely related to production in the previous year.           However, resource matching alone is inadequate to explain
These effects were independent of weather conditions,               the magnitude and regularity of variation in seed crops by
suggesting that intrinsic factors such as internal resource         many species (Kelly 1994; Koenig et al. 1994; Koenig and
budgets or an adaptive strategy of variable reproductive            Knops 2000). Masting may be an adaptive reproductive
output influenced fruit production. To our knowledge, this          strategy reflecting economies of scale if large, intermittent
is the first evidence of masting in members of the genera           reproductive efforts are more efficient than constant efforts
Cornus, Viburnum, and Amelanchier, and in members of                (Janzen 1978; Norton and Kelly 1988). For example,
Prunus and Sorbus in North America. All species produced            masting can enhance wind and insect pollination (Nilsson
fewer fruits when weather conditions were dry, so the               and Wästljung 1987; Crone and Lesica 2006), and seed
Moran effect could have synchronized fruit production               survival by satiating predators during mast years (Janzen
both within and among species. Patterns and causes of               1971; Silvertown 1980; Kelly et al. 2008a). Masting may
variation in berry crops have implications for ecosystem            be of no benefit or detrimental to dispersal in endozooch-
                                                                    orous species (those with fleshy fruits dispersed by mutu-
                                                                    alist frugivores) because dispersers would be saturated in
E. J. Howe (&)  M. E. Obbard  J. Bowman                           mast years (Janzen 1971; Silvertown 1980; Herrera et al.
Wildlife Research and Development Section,
                                                                    1998), unless dispersers were attracted to, or can show a
Ontario Ministry of Natural Resources, DNA Building,
Trent University, Peterborough, ON, K9J 7B8, Canada                 rapid numerical response to, large seed crops (Kelly 1994;
e-mail: eric.howe@ontario.ca                                        Vander Wall 2002). Indeed, reviews and meta-analyses

                                                                                                                      123
348                                                                                               Popul Ecol (2012) 54:347–356

have shown that masting is less common in endozoochor-          dispersed by passerine and gallinaceous birds (Stiles 1980;
ous species and shrubs than other plants (Silvertown 1980;      Willson 1986), and mammals including ungulates, rodents,
Herrera et al. 1998; Kelly and Sork 2002), though shrubs        and lagomorphs, but particularly carnivores such as racoons,
and North American endozoochorous species were under-           coyotes, red foxes, American martens, and American black
represented in the masting literature (Kelly and Sork 2002).    bears (Willson 1993). All of these potential dispersers were
Studies demonstrating negative impacts of masting on            common on our study area. We questioned whether the
dispersal of endozoochorous seeds have focused on avian         observed variation in seedfall was adequately explained by
frugivores (Davidar and Morton 1986; Herrera et al. 1994;       spatially correlated weather conditions, or if prior repro-
Herrera 1998). However, recent work from Spain suggests         ductive output had an effect independent of weather con-
that mammals, particularly carnivores, opportunistically        ditions. Independent effects of prior reproduction would
exploit large fruit crops and can be important dispersers of    suggest that intrinsic factors such as the internal resource
endozoochorous seeds (Martı́nez et al. 2008; Fedriani and       budgets of individual plants or an adaptive strategy of
Delibes 2009; Guitiàn and Munilla 2010; Matı́as et al.         variable reproductive output contributed to the observed
2010). Among North American carnivores, American black          pattern. We assessed synchrony of weather conditions and
bears (Ursus americanus), brown bears (U. arctos), rac-         seedfall data, and tested for effects of fruit production in
coons (Procyon lotor), coyotes (Canis latrans), red foxes       year t - 1, and soil moisture during fruiting in year t, on
(Vulpes vulpes), gray foxes (Urocyon cinereoargenteus),         fruit production by endozoochorous species representing
and American martens (Martes americana) are highly              eight genera. The study area covered more than 65000 km2
frugivorous and can be effective dispersers because they        within Canada’s boreal shield ecozone (Ecological Strati-
harvest fruits effectively, carry many seeds, may move          fication Working Group 1996; Fig. 1), and was largely
several kilometers before depositing seeds, and deposit         covered by forests and wetlands with tree and shrub species
viable seeds (Willson 1993; Borchert and Tyler 2010; also       typical of both boreal and eastern North American decidu-
see Koike et al. 2010).                                         ous forests (Rowe 1972), interspersed with open water and
    For any fitness benefits of masting to be realized, tem-    agricultural, rural, and urban development.
poral variation in seedfall must first become synchronized
among individuals. The Moran effect was suggested as a
likely synchronizing agent (Silvertown 1980). Masting           Methods
could then evolve via selection favouring individuals with
common responses to environmental cues and high inter-          Five Ontario Ministry of Natural Resources (OMNR)
annual variation in seedfall (Silvertown 1980; Kelly and        administrative districts monitored fruit and seed availabil-
Sork 2002). Theoretical work with resource budget models        ity for wildlife in central and eastern Ontario from
has shown that density-dependent pollen limitation could        1998–2009. Fruit or seed production by eight species, or
also synchronize intermittently reproducing individuals in      groups of species, of endozoochorous trees and shrubs
a population. This would result in population-level masting     (Table 1) was ranked on a five point scale in each district,
even in the absence of either environmental fluctuations or     where 0 represented crop failure and 4 represented a
fitness benefits, provided that most plants produced few        bumper crop. Sample sizes of observations varied among
seeds in some years (Isagi et al. 1997; Satake and Iwasa        districts and years. Not all districts completed surveys in all
2000). Empirical studies have shown that both the Moran         years, and no district recorded fruit production by all eight
effect and density-dependent pollen limitation can syn-         species in all 12 years. Furthermore, some district-specific
chronize interannual variation in seedfall by some species      rankings were consistently based on multiple observations
(Selås 2000; Rees et al. 2002; Schauber et al. 2002; Crone     from different locations (e.g., townships) within the
et al. 2005, 2009; Kon et al. 2005; Lamontagne and Boutin       district, but in other cases single values reflecting fruit
2007), and updated resource budget models demonstrate           production across entire districts were assigned by an
that both mechanisms acting in concert would produce            individual or by consensus among observers. Observers in
the strongest correlations in seedfall among individuals        some but not all districts visited the same sites each year,
(Satake and Iwasa 2002).                                        but fixed monitoring plots were not established. Despite
    Considerable annual variation in fruit production by        these differences in collection methods, the data showed
endozoochorous shrubs and small trees was apparent in           consistent temporal patterns in fruit production across the
seedfall data from central Ontario, Canada, 1998–2009. All      region (Ontario Ministry of Natural Resources, unpub-
the plant species we monitored are insect-pollinated, pri-      lished data; Obbard et al. 2003). Noyce and Coy (1990)
marily by bees, or in the case of Sambucus spp., primarily by   showed that a similar index of fruit production was con-
flies, and are also visited by other insects including wasps,   sistent with quantitative data and sufficient to demonstrate
beetles, butterflies, and moths. Seeds of these species are     differences among years and forest stands.

123
Popul Ecol (2012) 54:347–356                                                                                                          349

Fig. 1 Location of the study                                                   80°0 W
area (shaded grey) in central
Ontario, Canada. Boundaries                                                                       0    50   100          200 km
between Ontario Ministry of
Natural Resources
administrative districts are
shown as black lines within the
greater study area. Black circles                                                                     Quebec
are weather stations

                                                                                                                                    45°0 N

                                             Michigan

                                                                             Ontario

                                                                                                 New York

Table 1 Plant species or
                                    Common name                  Family         Genus         Species included
groups of species monitored in
central Ontario, Canada,            Mountain ash                 Rosaceae       Sorbus        S. americana, S. decora
1998–2009
                                    Cherry                       Rosaceae       Prunus        P. serotina, P. pensylvanica, P. virginiana
                                    Juneberry (serviceberry/     Rosaceae       Amelanchier   A. alnifolia, A. Canadensis, A. arborea,
                                      saskatoonberry/shadbush)                                 A. laevis, A. bartramiana
                                    Red raspberry                Rosaceae       Rubus         R. ideaus
                                    Nannyberry                   Adoxaceae      Viburnum      V. lentago
                                    Elderberry                   Adoxaceae      Sambucus      S. canadensis, S. pubens
                                    Dogwoods                     Cornaceae      Cornus        C. stolonifera, C. alternifolia, C. rugosa,
                                                                                               C. racemosa
                                    Blueberry                    Ericaceae      Vaccinium     V. angustifolium, V. myrtilloides

   To assess synchrony in seed production among districts,            evapotranspiration data over the previous 52 days (Turner
we calculated Buonaccorsi et al.’s (2001) measure indi-               1972; Van Wagner 1987). The DC therefore combined
cating whether pairs of time series from different districts          weather variables to provide a measure of the amount of
changed in agreement (A) and average A across pairs of                soil moisture available to plants. We selected DC values
         Because district-specific time series had missing
series (A).                                                           on 15 August each year from the ten weather stations in
values and one missing value in a time series reduced the             our study area for which continuous data were available
number of possible comparisons between any two districts              (Fig. 1). We tested for synchrony in DC data using the
across consecutive years by two, sample sizes for some                modified correlogram procedure (Koenig and Knops
measures of agreement were small. We did not calculate A              1998) with 1000 trials. We were more interested in
where there were fewer than 4 possible comparisons within             whether DC varied synchronously within the entire study
the 11 year time series, or A where there were less than 3           area than the decline in synchrony with distance within
pairwise comparisons between Districts with adequate data             the study area, so we applied the test to the complete data
to calculate A.                                                       set, and to two subcategories based on distance between
   A drought code (DC) was calculated as part of                      weather stations. Means and variances of station-specific
Ontario’s Forest Fire Weather Index. It described the                 DC data were not correlated with each other (R2 \ 0.10),
dryness of the organic soil layer below the fine fuel and             and no trends were apparent, so the tests were applied to
duff layers and was calculated from daily rainfall and                untransformed data.

                                                                                                                                  123
350                                                                                                  Popul Ecol (2012) 54:347–356

    We averaged seed production scores across districts and            The mean (SD) number of OMNR Districts estimating
DC values across weather stations before fitting regression         fruit production annually was 4.2 (0.80), and the mean
models to test for effects of weather and prior reproduction        annual number of observations was 31 (6.5). The number
on fruit production by each species or group of species. We         of Districts reporting species-specific fruit production
first averaged fruit production scores across multiple              annually averaged 4.1 (0.79) for mountain ash, 3.3 (0.75)
reports from within each district and year (where available)        for dogwood, 3.7 (0.78) for juneberry, 2.4 (1.17) for nan-
to give equal weight to data from each district. We then            nyberry, 4.2 (0.84) for cherry, 3.9 (1.00) for elderberry and
averaged scores across districts in each year to obtain             blueberry, and 4.1 (0.80) for raspberry. Fruit production by
regional time series for each species. By averaging fruit           all species for which we calculated A tended to increase or
production scores, we avoided missing values in the time            decrease relative to the previous year synchronously across
series; however, we also removed any spatial variation in           the study area (Table 2). Mountain ash and dogwood
seed production among districts and were required to                always fluctuated in the same direction across all pairs of
assume that seed production scores from a subset of dis-            districts for which data were available. Relative fruit pro-
tricts was representative of production across the entire           duction in consecutive years by other species agreed in
study area in some years. We calculated A (Buonaccorsi              62–78% of cases (Table 2). Mean fruit production by
et al. 2001) between regional average fruit production              mountain ash, nannyberry, juneberry, dogwood, cherry,
scores for each pair of species to assess whether fruit             and elderberry fluctuated synchronously across the region
production varied synchronously among species across the            (Table 3). Fluctuations in fruit production by blueberry
region.                                                             were in the same direction as fluctuations by other species
    We duplicated fruit production data for each species or         55% of the time, and fluctuations in fruit production by
group of species at a 1-year lag. Prior to fitting regression       raspberry were in the same direction as fluctuations by
models, we examined Pearson correlation coefficients                other species 55–73% of the time.
(rp; 1) among fruit production scores of the different species or      Fruit production by some species, such as raspberry,
groups of species, (2) between production by each species           appeared to vary inversely with DC, whereas for other
or group of species in year t (prodt) and in year t - 1             species, such as mountain ash, a pattern of good and poor
(prodt-1), (3) between prodt and DC, and (4) to assess              fruit production in alternating years was apparent (Fig. 2).
collinearity between predictors, between prodt-1 and DC.            Poor fruit production by species that exhibited a biennial
The same set of four linear regression models was then              pattern still coincided with dry conditions, which were
fitted to regional data (1999–2009) for each species or             most severe in 2001, 2005, and 2007 (Fig. 2).
group of species to test for effects of fruit production in            Correlation coefficients between prodt for different
year t - 1 and summer weather conditions in year t on fruit         species were all positive; those between prodt and DC were
production in year t. Regression models were a null                 all negative, as were those between prodt and prodt-1
(intercept only) model, models with effects of each of              except in the case of raspberry (Table 4). Collinearity
prodt-1 and DC, and a model with additive effects of both           between DC and prodt-1 as predictors of prodt was not a
predictors. We compared models using AICc (Hurvitch and             major concern rp \ 0.3; Table 4). Distributions of regional
Tsai 1989). We hierarchically partitioned coefficients of           fruit production scores and DC values did not differ from
multiple determination (R2) from the additive model into            normal (Shapiro–Wilk’s W [ 0.88; P [ 0.10 in all cases).
independent and joint contributions by each predictor               Nonlinearity of relationships was not apparent in plots of
(Chevan and Sutherland 1991). Statistical analyses were             prodt versus DC and prodt-1. Relative AICc values among
performed using R software (R Development Core Team                 the four models fit to data for each species indicated that
2009).                                                              fruit production by mountain ash, dogwood, juneberry,
                                                                    nannyberry, and possibly cherry was affected by fruit
                                                                    production in the previous year, and that fruit production
Results                                                             by juneberry, nannyberry, cherry, elderberry, blueberry,
                                                                    and raspberry was affected by dry summer weather con-
DC values varied synchronously across the study area                ditions (Table 5). Hierarchical partitioning further revealed
(mean rp across all pairs of weather stations 0.66, SD 0.17,        that prior reproduction accounted for more of the explained
P = 0.002 by the modified correlogram procedure applied             variation in fruit production by mountain ash, dogwood,
to the entire data set). Synchrony declined slightly with           and juneberry, whereas weather conditions accounted for
distance (mean rp across weather stations an average of 111         more of the explained variation in fruit production by
and 263 km apart = 0.70 and 0.61, respectively), but was            cherry, elderberry, blueberry, and raspberry; contributions
significant (P \ 0.05) at both distance categories.                 to nannyberry production were similar (Table 5). Joint

123
Popul Ecol (2012) 54:347–356                                                                                                                            351

Table 2 Agreement in the
                                                        Species                 A              Mean # comparisons               Number of pairwise A values
direction of change in fruit
                                                                                                contributing to pairwise         contributing to A
production scores among
                                                                                                A (maximum = 10                  (maximum = 10
administrative districts in
                                                                                                time intervals)                  pairs of Districts)
central Ontario, Canada,
1998–2009                                               Mountain ash            1.00            6.3                              7
                                                        Dogwood                 1.00            5.0                              3
                                                        Juneberry               0.72            6.0                              3
                                                        Nannyberry              n/a             5.0                              1
A measures the mean proportion                         Cherry                  0.78            6.3                              8
of time intervals during which                          Elderberry              0.62            6.2                              5
pairs of series agreed in their                         Blueberry               0.67            5.4                              5
change of direction
                                                        Raspberry               0.68            6.1                              7
(Buonaccorsi et al. 2001)

Table 3 Proportions of years
                                                                       NN             JN          DW             CH         EL            BB            RA
with the same direction of
change in the amount of fruit                           MA             1.00           1.00        1.00           1.00       1.00          0.55          0.73
produced by pairs of plant
species in central Ontario,                             NN                            1.00        1.00           1.00       1.00          0.55          0.64
Canada, 1998–2009                                       JN                                        1.00           1.00       1.00          0.55          0.64
MA mountain ash,                                        DW                                                       1.00       1.00          0.55          0.73
NN nannyberry, JNjuneberry,                             CH                                                                  1.00          0.55          0.64
DW dogwood, CH cherry,                                  EL                                                                                0.55          0.64
EL elderberry, BB blueberry,
                                                        BB                                                                                              0.55
RA raspberry

                                                                                             between standardized residuals and theoretical quantiles
                                    4.0
 Seed production and drought code

                                                                                             were approximately linear suggesting errors were normally
                                    3.5                                                      distributed.

                                    3.0

                                    2.5
                                                                                             Discussion

                                    2.0                                                      Our data describe a system where weather conditions
                                                                                             varied synchronously, and fruit production by several
                                    1.5
                                                                                             species of endozoochorous shrubs and small trees appeared
                                    1.0                                                      to fluctuate synchronously, over more than 65000 km2 of
                                                                                             boreal-deciduous transitional forests in eastern North
                                          2000   2002   2004         2006     2008
                                                                                             America. To our knowledge, this is the first evidence of
                                                        Year
                                                                                             masting in the genera Cornus, Viburnum, and Amelanchier,
Fig. 2 Drought code values (circles, solid line) on 15 August, and                           and in members of Prunus and Sorbus in North America.
fruit production scores for mountain ash (grey triangles and grey                            Fruit production was low when weather conditions during
dashed line), raspberry (grey squares and grey dotted line), cherry
                                                                                             fruiting were dry, but variation in fruit production by
(white triangles and black dashed line), and nannyberry (white
squares and black dotted line) in central Ontario, Canada, 1999–2009.                        mountain ash, dogwood, juneberry, nannyberry, and
Drought code values were scaled by dividing by 100                                           possibly cherry was better explained by inverse relationships
                                                                                             between fruit production in the current and previous year,
contributions were consistently smaller than independent                                     or a combination of such relationships and weather con-
contributions (Table 5), so high R2 values were not                                          ditions, than by weather conditions alone. We conclude
attributable to collinearity between predictors (Chevan and                                  that the observed interannual variation in fruit production
Sutherland 1991). Plots of residuals versus fitted values                                    was not solely a result of a Moran effect, and that other
from the most general model fit to data for each species did                                 factors such as an evolved strategy of variable reproductive
not suggest violations of the assumptions of homoscedas-                                     output (Kelly 1994), or the energetic costs of reproduction
ticity of errors or linearity of relationships. Relationships                                in combination with density-dependent pollen limitation

                                                                                                                                                 123
352

123
      Table 4 Pearson correlation coefficients between fruit production scores for different species or groups of species, between productivity scores in year t and in year t - 1, and between drought
      code values on 15 August (DC) and fruit production scores in the same and in previous years in central Ontario, Canada, 1999–2009
                 DCA       MA         NN         JN         DW         CH         EL         BB         RA         MA-1       NN-1       JN-1       DW-1       CH-1       EL-1       BB-1       RA-1

      DCA        1         -0.46      -0.68      -0.44      -0.52      -0.72      -0.7       -0.65      -0.61        0.26       0.15       0.08       0.27       0.14       0.06       0.02       0
      MA                     1          0.86       0.83       0.96       0.85       0.83       0.38       0.22     -0.79      -0.76      -0.83      -0.77      -0.81      -0.55      -0.18      -0.46
      NN                                1          0.8        0.9        0.85       0.93       0.37       0.17     -0.76      -0.69      -0.65      -0.66      -0.6       -0.47      -0.18      -0.39
      JN                                           1          0.86       0.8        0.73       0.25       0.12     -0.74      -0.79      -0.79      -0.62      -0.72      -0.61      -0.01      -0.68
      DW                                                      1          0.92       0.85       0.38       0.18     -0.76      -0.75      -0.83      -0.69      -0.71      -0.5       -0.11      -0.38
      CH                                                                 1          0.84       0.42       0.47     -0.59      -0.52      -0.62      -0.52      -0.52      -0.26        0.07     -0.29
      EL                                                                            1          0.5        0.34     -0.59      -0.6       -0.49      -0.53      -0.51      -0.33      -0.1       -0.29
      BB                                                                                       1          0.37       0.07     -0.07        0        -0.07      -0.05        0        -0.31        0.06
      RA                                                                                                  1          0.15       0.25       0.11       0.1        0.13       0.41       0.55       0.17
      MA-1                                                                                                           1          0.89       0.86       0.94       0.9        0.83       0.24       0.45
      NN-1                                                                                                                      1          0.86       0.87       0.88       0.9        0.21       0.43
      JN-1                                                                                                                                 1          0.79       0.83       0.7        0.05       0.46
      DW-1                                                                                                                                            1          0.94       0.86       0.32       0.29
      CH-1                                                                                                                                                       1          0.88       0.36       0.53
      EL-1                                                                                                                                                                  1          0.42       0.45
      BB-1                                                                                                                                                                             1          0.27
      RA-1                                                                                                                                                                                        1
      MA mountain ash, NN nannyberry, JN juneberry, DW dogwood, CH cherry, EL elderberry, BB blueberry, RA raspberry
                                                                                                                                                                                                          Popul Ecol (2012) 54:347–356
Popul Ecol (2012) 54:347–356                                                                                                                 353

Table 5 Model comparisons and hierarchical partitioning of R2 values from linear regression models fit to fruit production data for different
species or groups of species in central and eastern Ontario, Canada 1999–2008
                  Model selection                                                           Partitioning R2
                  Null        prodt-1             DC                   prodt-1 ? DC         Contribution to R2             Ratio of independent
                                                                                                                           contributions to R2
                  DAICc       DAICc      R2       DAICc       R2       DAICc       R2       prodt-1     DC        Joint    %(prodt-1): %(DC)

Mountain ash      6.95        0.00       0.63     8.23        0.21     2.98        0.68     0.56        0.14      0.07     80:20
Dogwood           3.23        0.00       0.48     3.69        0.27     2.36        0.60     0.40        0.20      0.08     67:33
Juneberry         7.21        0.25       0.63     8.74        0.20     0.00        0.77     0.60        0.17      0.02     78:22
Nannyberry        9.66        6.39       0.48     6.71        0.47     0.00        0.82     0.42        0.40      0.06     51:49
Cherry            4.04        4.44       0.27     0.00        0.52     0.03        0.70     0.23        0.47      0.05     33:67
Elderberry        3.60        6.28       0.11     0.00        0.50     3.26        0.58     0.09        0.48      0.01     16:84
Blueberry         2.16        4.96       0.10     0.00        0.43     3.33        0.52     0.09        0.42      0.00     18:82
Raspberry         1.11        4.71       0.03     0.00        0.37     4.69        0.40     0.03        0.37      0.00     8:92
Shown are DAICc values among models fit to the same data (DAICc B1 appear in bold), coefficients of determination (R2) for each model,
independent contributions to multiple R2 by each of productivity in the previous year (prodt-1) and the drought code on 15 August (DC), the joint
contribution to R2, and the ratio of independent contributions to R2 as a percent of the total explained variation

(Isagi et al. 1997) contributed to the observed variation.                apparently causing the synchrony among species (Laine
Our results support previous research that showed the                     and Henttonen 1983; Selås 1997). Here, the DC was sup-
Moran effect to be insufficient to explain masting (Kelly                 ported as a predictor of fruit production in six of the eight
1994; Koenig et al. 1994; Koenig and Knops 2000), but                     species monitored, and production by the other two species
they do not allow us to infer whether pollen limitation                   was also poor in dry years. Spatially correlated weather
affected fruit production, or whether masting was adaptive.               conditions may therefore have synchronized interannual
Pollen limitation is widespread in angiosperms but there                  variation in fruit production both within and among species
have been few studies of it in the species we monitored                   on our study area. Other studies identified positive rela-
(Burd 1994; Ashman et al. 2004; Knight et al. 2005).                      tionships between fruit production and precipitation in the
Enhanced seed survival through predator satiation is a                    same year (Selås 2000; Selås et al. 2001; Krebs et al.
common benefit of masting in animal-pollinated plants                     2009).
(Sork 1993; Kelly and Sork 2002), and was apparent in                        Density-dependent pollen limitation might also have
studies of Sorbus and Prunus elsewhere (Kobro et al. 2003;                contributed to the intraspecific synchrony we observed
Li and Zhang 2007), but we are not aware of studies of pre-               (Isagi et al. 1997; Rees et al. 2002; Crone et al. 2005,
dispersal seed predation of the species we monitored.                     2009). Furthermore, since the species we monitored share
Mammalian carnivores common on our study area could                       pollinators, release from pollen limitation could promote
satisfy the conditions necessary for masting to enhance                   synchrony among species if pollinator populations were
dispersal, in that they are attracted to large fruit crops, may           attracted to, or responded numerically to, large flowering
not be satiated in mast years, and effectively disperse                   efforts (Satake and Iwasa 2002; Tachiki et al. 2010). If
endozoochorous seeds (Willson 1993; Borchert and Tyler                    density-dependent pollen limitation is necessary to syn-
2010; Guitiàn and Munilla 2010).                                         chronize individuals (Satake and Iwasa 2002), spatially
   Fruit production by mountain ash, dogwood, nannyberry,                 correlated weather conditions such as droughts or spring
juneberry, cherry, and elderberry always fluctuated in                    frosts that kill flower buds (e.g., Usui et al. 2005) could
the same direction between years, and blueberry and                       provide the years of low population-level fruiting success
raspberry production fluctuated in the same direction as                  which are required for synchrony to occur under these
other species more than 50% of the time. Annual variation                 models (Isagi et al. 1997; Satake and Iwasa 2000).
in fruit production by different endozoochorous species                      Pulses of food resources produced in mast years cascade
including several of the ones we monitored was also syn-                  through ecosystems (Ostfeld and Keesing 2000; Liebhold
chronous over 20 years in a similar ecosystem in Minnesota                et al. 2004) and effects on higher trophic levels are well-
(Garshelis and Noyce 2008); berry crops in southwest Yukon                documented (King 1983; Satake et al. 2004; Kelly et al.
tended to covary among species as well (Krebs et al. 2010).               2008b; Schmidt and Ostfeld 2008). Most studies demon-
Flower, berry, and seed production by several plants includ-              strating effects of variable seedfall on North American seed
ing Vaccinium myrtillus and Empetrum nigrum varied                        consumers focused on responses to seedfall from trees
synchronously in Fennoscandia, with summer precipitation                  (Eiler et al. 1989; Koenig and Knops 2001; Bowman et al.

                                                                                                                                     123
354                                                                                                              Popul Ecol (2012) 54:347–356

2005, 2008; Falls et al. 2007), but Kelly et al. (2008b)                 References
showed that the effects of masting on other trophic levels
vary among ecosystems. Effects of masting by endozo-                     Ashman T-L, Knight TM, Steets JA, Amarasekare P, Burd M,
                                                                              Campbell DR, Dudash MR, Johnston MO, Mazer SJ, Mitchell
ochorous plants may be more apparent in boreal forests                        RJ, Morgan MT, Wilson WG (2004) Pollen limitation of plant
where the relative importance of berries to wildlife is                       reproduction: ecological and evolutionary causes and conse-
greater. For example, in Fennoscandia, abundance of                           quences. Ecology 85:2408–2421
grouse, voles, and raptors, and mass of moose calves cor-                Borchert M, Tyler CM (2010) Desiccation sensitivity and heat
                                                                              tolerance of Prunus ilicifolia seeds dispersed by American black
related with annual variation in fruit production by bilberry                 bears (Ursus americanus). West N Am Nat 70:457–466
(Vaccinium myrtillus) and cowberry (V. vitis-idaea; Laine                Bowman J, Holloway GL, Malcom JR, Middel KR, Wilson PJ (2005)
and Henttonen 1983; Selås 1997; Selås et al. 2001). In the                  Northern range boundary dynamics of southern flying squirrels:
southwestern Yukon, Canada, autumn abundance of the                           evidence of an energetic bottleneck. Can J Zool 83:1486–1494
                                                                         Bowman J, Phoenex RD, Sugar A, Dawson FN, Holborn G (2008)
four dominant rodent species was explained well by vari-                      Spatial and temporal dynamics of small mammals at a regional
ation in berry crops (Krebs et al. 2010), and in the boreal                   scale in Canadian boreal forest. J Mammal 89:381–387
forest of Ontario, Canada, population-level reproductive                 Buonaccorsi JP, Elkinton JS, Evans SR, Liebhold AM (2001)
failures by black bears occurred when berry crops failed                      Measuring and testing for spatial synchrony. Ecology 82:1668–
                                                                              1679
(Obbard and Howe 2008).                                                  Burd M (1994) Bateman’s principal and plant reproduction: the role
   Data limitations reduce the strength of our inferences.                    of pollen limitation in fruit set and seed set. Bot Rev 60:83–139
We analyzed rank data collected by different observers                   Chevan A, Sutherland M (1991) Hierarchical partitioning. Am Stat
with little standardization of data collection methods, and                   45:90–96
                                                                         Crone EE, Lesica P (2006) Pollen and water limitation in Astragalus
fruit production data were specific to a group of species                     scaphoides, a plant that flowers in alternate years. Oecologia
rather than a single species in six of eight cases which                      150:40–49
could have obscured species-specific patterns. We tested                 Crone EE, Polansky L, Lesica P (2005) Empirical models of pollen
for negative autocorrelations between current and prior                       limitation, resource acquisition, and mast seeding by a bee-
                                                                              pollinated wildflower. Am Nat 166:396–408
reproduction at a 1-year lag only but plant populations                  Crone EE, Miller E, Sala A (2009) How do plants know when other
could have masted at longer or variable intervals (e.g.,                      plants are flowering? Resource depletion, pollen limitation and
Selås 1997; Satake and Bjørnstad 2008), and we did not                       mast-seeding in a perennial wildflower. Ecol Lett 12:1119–1126
test for effects of all weather variables that might have                Davidar P, Morton ES (1986) The relationship between fruit crop
                                                                              sizes and fruit removal rates by birds. Ecology 67:262–265
affected fruit production. Synchrony in fruit production                 Ecological Stratification Working Group (1996) A National Ecolog-
among districts, though apparent, could not be demon-                         ical Framework for Canada. Agriculture and Agri-Food Canada,
strated statistically. Asynchrony among districts would                       Research Branch, Centre for Land and Biological Resources
dampen interannual variation in regional average rankings,                    Research and Environment Canada, State of Environment
                                                                              Directorate, Ottawa/Hull, Canada
potentially increasing the risk of type II error in tests for            Eiler JH, Wathen WG, Pelton MR (1989) Reproduction in black
effects of prior reproduction. Despite these limitations we                   bears in the southern Appalachian mountains. J Wildl Manag
think that our novel findings raise interesting questions that                53:353–360
will encourage future studies of this subject.                           Falls JB, Falls EA, Fryxell JM (2007) Fluctuations of deer mice in
                                                                              Ontario in relation to seed crops. Ecol Monogr 77:19–32
   Our results suggest testable hypotheses about the con-                Fedriani JM, Delibes M (2009) Seed dispersal in the Iberian pear,
tribution of adaptive reproductive strategies and pollen                      Pyrus bourgaeana: a role for infrequent mutualists. Ecoscience
limitation to masting in endozoochorous plants. Studies                       16:311–321
could avoid the limitations of our dataset by collecting                 Garshelis DL, Noyce KV (2008) Seeing the world through the nose of
                                                                              a bear—diversity of foods fosters behavioral and demographic
quantitative seedfall data at fixed locations over a longer                   stability. In: Fullbright TE, Hewitt DG (eds) Wildlife science.
time period while monitoring seed survival and responses                      Linking ecological theory and management applications. CRC
of pollinators and consumers. In particular, knowledge of                     Press, Boca Raton
rates of pre-dispersal seed predation and the role of                    Guitiàn J, Munilla I (2010) Responses of mammal dispersers to fruit
                                                                              availability: Rowan (Sorbus aucuparia) and carnivores in moun-
mammals as dispersers might suggest adaptive benefits,                        tain habitats of northern Spain. Acta Oecol 16:242–247
and manipulative experiments such as those conducted by                  Herrera CM (1998) Long-term dynamics of Mediterranean frugivo-
Crone et al. (2005, 2009) could be used to test for effects of                rous birds and fleshy fruits: a 12-yr study. Ecol Monogr 68:511–
resource accumulation and pollen limitation.                                  538
                                                                         Herrera CM, Jordano P, López-Soria L, Amat JA (1994) Recruitment
                                                                              of a mast-fruiting, bird-dispersed tree: bridging frugivore activity
Acknowledgments We thank J. Caputo, Aviation, Forest Fire and
                                                                              and seedling establishment. Ecol Monogr 64:315–344
Emergency Management Branch, OMNR for providing weather data,
                                                                         Herrera CM, Jordano P, Guitián J, Traveset A (1998) Annual
the many OMNR district biologists and technicians who collected and
                                                                              variability in seed production by woody plants and the masting
maintained fruit production data, and A. Foley for discussing patterns
                                                                              concept: reassessment of principals and relationship to pollina-
in annual numbers of seeds collected for the Ontario Tree Seed
                                                                              tion and dispersal. Am Nat 152:576–594
Facility.

123
Popul Ecol (2012) 54:347–356                                                                                                                  355

Hurvitch CM, Tsai C-L (1989) Regression and time series model             Martı́nez I, Garcı́a D, Obeso JR (2008) Differential seed dispersal
     selection in small samples. Biometrika 76:297–307                          patterns generated by a common assemblage of vertebrate
Isagi Y, Sugimura K, Sumida A, Ito H (1997) How does masting                    frugivores in three fleshy-fruited trees. Ecoscience 15:189–199
     happen and synchronize? J Theor Biol 187:231–239                     Matı́as L, Zamora R, Mendoza I, Hódar JA (2010) Seed dispersal
Janzen DH (1971) Seed predation by animals. Annu Rev Ecol Syst                  patterns by large frugivorous mammals in a degraded mosaic
     2:465–492                                                                  landscape. Restor Ecol 18:619–627
Janzen DH (1978) Seeding patterns of tropical trees. In: Tomlinson        Nilsson SG, Wästljung U (1987) Seed predation and cross-pollination
     PB, Zimmermann MH (eds) Tropical trees as living systems.                  in mast-seeding beech (Fagus sylvatica) patches. Ecology
     Cambridge University Press, Cambridge                                      68:260–265
Kelly D (1994) The evolutionary ecology of mast seeding. TREE             Norton DA, Kelly D (1988) Mast seeding over 33 years by
     9:465–470                                                                  Dacrydium cupressinum Lamb. (rimu) (Podocarpaceae) in New
Kelly D, Sork VL (2002) Mast seeding in perennial plants: why, how,             Zealand: the importance of economies of scale. Funct Ecol
     where? Annu Rev Ecol Syst 33:427–447                                       2:399–408
Kelly D, Turnbull MH, Pharis RP, Sarfati MS (2008a) Mast seeding          Noyce KV, Coy PL (1990) Abundance and productivity of bear food
     predator satiation, and temperature cues in Chionochloa                    species in different forest types of northcentral Minnesota. Int
     (Poaceae). Popul Ecol 50:343–355                                           Conf Bear Res Manag 8:169–181
Kelly D, Koenig WD, Liebhold AM (2008b) An intercontinental               Obbard ME, Howe EJ (2008) Demography of black bears in hunted
     comparison of the dynamic behavior of mast seeding commu-                  and unhunted areas of the boreal forest of Ontario. J Wildl
     nities. Popul Ecol 50:329–342                                              Manag 72:869–880
King CM (1983) The relationships between beech (Nothofagus spp.)          Obbard ME, Pond BA, Howe EJ (2003) Analysis of relationships
     seedfall and populations of mice (Mus musculus), and the                   among black bear nuisance activity, food availability, and
     demographic and dietary response of stoats (Mustela erminea),              harvest in Ontario. Appendix 10. In: Poulin R, Knight J, Obbard
     in three New Zealand forests. J Anim Ecol 52:141–166                       M, Whitherspoon G (eds) Nuisance bear review committee
Knight TM, Steets JA, Vamosi JC, Mazer SJ, Burd M, Campbell DR,                 report and recommendations. Ontario Ministry of Natural
     Dudash MR, Johnston MO, Mitchell RJ, Ashman T-L (2005)                     Resources, Peterborough
     Pollen limitation of plant reproduction: pattern and process.        Ostfeld RS, Keesing F (2000) Pulsed resources and consumer commu-
     Annu Rev Ecol Evol Syst 36:467–497                                         nity dynamics in terrestrial ecosystems. TREE 15:232–237
Kobro S, Søreide L, Djønne E, Rafoss T, Jaastad G, Witzgall P (2003)      R Development Core Team (2009) R: a language and environment for
     Masting of rowan Sorbus aucuparia L. and consequences for the              statistical computing. R Foundation for Statistical Computing,
     apple fruit moth Argyresthia conjugella Zeller. Popul Ecol                 Vienna, Austria. ISBN 3-900051-07-0. http://www.R-project.org
     45:25–30                                                             Rees M, Kelly D, Bjørnstad ON (2002) Snow tussocks, chaos, and the
Koenig WD, Knops MH (1998) Testing for spatial autocorrelation in               evolution of mast seeding. Am Nat 160:44–59
     ecological studies. Ecography 21:423–429                             Rowe JS (1972) Forest Regions of Canada. Forest Service Publication
Koenig WD, Knops MH (2000) Patterns of annual seed production                   No. 1300. Department of Fisheries and the Environment, Ottawa,
     by northern hemisphere trees: a global perspective. Am Nat                 Ontario, Canada
     155:59–69                                                            Royama T (1992) Analytical population dynamics. Chapman and
Koenig WD, Knops MH (2001) Seed-crop size and eruptions of North                Hall, London
     American boreal seed-eating birds. J Anim Ecol 70:609–620            Satake A, Bjørnstad ON (2008) A resource budget model to explain
Koenig WD, Mumme RL, Carmen WJ, Stanback MT (1994) Acorn                        intraspecific variation in mast reproductive dynamics. Ecol Res
     production by oaks in central coastal California: variation within         23:3–10
     and among years. Ecology 75:99–109                                   Satake A, Iwasa Y (2000) Pollen coupling of forest trees: forming
Koike S, Masaki T, Nemoto Y, Kozakai C, Yamazaki K, Kasai S,                    synchronized and periodic reproduction out of chaos. J Theor
     Nakajima A, Kaji K (2010) Estimate of the seed shadow created              Biol 203:63–84
     by the Asiatic black bear Ursus thibetanus and its characteristics   Satake A, Iwasa Y (2002) The synchronized and intermittent
     as a seed disperser in Japanese cool-temperate forest. Oikos               reproduction of forest trees is mediated by the Moran effect,
     120:280–290                                                                only in association with pollen coupling. J Ecol 90:830–838
Kon H, Noda T, Terazawa K, Koyama H, Yasaka M (2005)                      Satake A, Bjørnstad ON, Kobro S (2004) Masting and trophic
     Proximate factors causing mast seeding in Fagus crenata:                   cascades: interplay between rowan trees, apple fruit moth, and
     the effects of resource level and weather cues. Can J Bot                  their parasitoid in southern Norway. Oikos 104:540–550
     83:1402–1409                                                         Schauber EM, Kelly D, Turchin P, Simon C, Lee WG, Allen RB,
Krebs CJ, Boonstra R, Cowcill K, Kenny AJ (2009) Climatic                       Payton IJ, Wilson PR, Cowan PE, Brokie RE (2002) Masting by
     determinants of berry crops in the boreal forest of the                    eighteen New Zealand plant species: the role of temperature as a
     southwestern Yukon. Botany 87:401–408                                      synchronizing cue. Ecology 83:1214–1225
Krebs CJ, Cowcill K, Boonstra R, Kenny AJ (2010) Do changes in            Schmidt KA, Ostfeld RS (2008) Numerical and behavioral effects
     berry crops drive population fluctuations in small rodents in the          within a pulse-driven system: consequences for shared prey.
     southwestern Yukon? J Mammal 91:500–509                                    Ecology 89:635–646
Laine K, Henttonen H (1983) The role of plant production in               Selås V (1997) Cyclic population fluctuations of herbivores as an
     microtine cycles in northern Fennoscandia. Oikos 40:407–418                effect of cyclic seed cropping of plants: the mast depression
Lamontagne JM, Boutin S (2007) Local-scale synchrony and                        hypothesis. Oikos 80:257–268
     variability in mast seed production patterns of Picea glauca.        Selås V (2000) Seed production of a masting dwarf shrub, Vaccinium
     J Ecol 95:991–1000                                                         myrtillus, in relation to previous reproduction and weather. Can J
Li H, Zhang Z (2007) Effects of mast seeding and rodent abundance               Bot 78:423–429
     on seed predation and dispersal by rodents in Prunus armeniaca       Selås V, Sonerud GA, Histøl T, Hjeljord O (2001) Synchrony in
     (Rosaceae). For Ecol Manag 242:511–517                                     short-term fluctuations of moose calf body mass and bank vole
Liebhold A, Koenig WD, Bjørnstad ON (2004) Spatial synchrony in                 population density supports the mast depression hypothesis.
     population dynamics. Annu Rev Ecol Syst 35:467–490                         Oikos 92:271–278

                                                                                                                                     123
356                                                                                                            Popul Ecol (2012) 54:347–356

Silvertown JW (1980) The evolutionary ecology of mast seeding in        Usui M, Kevan PG, Obbard M (2005) Pollination and breeding
     trees. Biol J Lin Soc 14:235–250                                        system of lowbush blueberries, Vaccinium angustifolium Ait,
Sork VL (1993) Evolutionary ecology of mast-seeding in temperate             and V. myrtilloides Michx. (Ericacaeae), in the boreal forest. Can
     and tropical oaks (Quercus spp.). Vegetation 107(108):133–147           Field Nat 19:48–57
Stiles EW (1980) Patterns of fruit presentation and seed dispersal in   Van Wagner CE (1987) Development and structure of the Canadian
     bird-disseminated woody plants in the eastern deciduous forest.         Forest Fire Weather Index System. Canadian Forest Service
     Am Nat 116:670–688                                                      Forestry Technical Report 35, Ottawa, Ontario, Canada
Tachiki Y, Iwasa Y, Satake A (2010) Pollinator coupling can induce      Vander Wall SB (2002) Masting in animal-dispersed pines facilitates
     synchronized flowering in different plant species. J Theor Biol         seedling dispersal. Ecology 83:3508–3816
     267:153–163                                                        Willson MF (1986) Avian frugivory and seed dispersal in eastern
Turner JA (1972) The Drought Code component of the Canadian                  North America. Curr Ornithol 3:223–279
     Forest Fire Behaviour System. Canadian Forest Service Publi-       Willson MF (1993) Mammals as seed-dispersal mutualists in North
     cation 1316, Ottawa, Ontario, Canada                                    America. Oikos 67:159–176

123
You can also read